检索结果-内蒙古大学图书馆

BSM-NET: multi-bandwidth, multi-scale and multi-modal fusion network for 3D object detection of 4D radar and LiDAR

MEASUREMENT SCIENCE AND TECHNOLOGY 2025年第3期36卷 036107-036107页

作者： Jiang, Tiezhen Kang, Runjie Li, Qingzhu Anhui Univ Sch Elect & Informat Engn Hefei 230601 Peoples R China

In recent years, with the rapid advancement of autonomous driving technology, the requirements for environmental perception tasks have become increasingly important. Four-dimensional (4D) millimeter-wave radar, an economical and reliable technology, has begun to attract attention. Additionally, LiDAR can measure accurately and is not easily interfered with, so it is also very popular. To overcome the limitations of single-sensor perception, this paper proposes the BSM-NET method, a multi-bandwidth, multi-scale, multi-modal fusion approach for 4D radar and LiDAR. This paper uses image technology to clean point cloud data, reducing errors and noise. BSM-NET consists of two key modules: multi-bandwidth fusion (MBF) and multi-scale fusion (MSF). MBF enhances data quality by capturing point cloud density and addressing issues such as gap filling and noise reduction. MSF improves accuracy and robustness through high-precision calculations. For better integration, we use the radar and LiDAR based multi-modal fusion (RLMF), which enables different types of data to learn from each other, allowing us to effectively fuse 80-line LiDAR and 4D radar data. Experimental results demonstrate that BSM-NET significantly outperforms the current state-of-the-art algorithms on Dual-Radar dataset. Compared to M2-Fusion, our fusion method achieves notable performance improvements of 2.04% and 2.21% in medium-difficulty 3D object detection and bird's eye view detection for the vehicle category, respectively.

关键词： multimodal fusion object detection autonomous driving 4D radar

来源：评论

学校读者我要写书评

暂无评论

Divide-and-Conquer: Confluent Triple-Flow Network for RGB-T Salient object detection

引用

IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE 2025年第3期47卷 1958-1974页

作者： Tang, Hao Li, Zechao Zhang, Dong He, Shengfeng Tang, Jinhui Nanjing Univ Sci & Technol Sch Comp Sci & Engn Nanjing 210094 Peoples R China Hong Kong Univ Sci & Technol Dept Elect & Comp Engn Hong Kong Peoples R China Singapore Management Univ Sch Comp & Informat Syst Singapore 188065 Singapore

RGB-Thermal Salient object detection (RGB-T SOD) aims to pinpoint prominent objects within aligned pairs of visible and thermal infrared images. A key challenge lies in bridging the inherent disparities between RGB and Thermal modalities for effective saliency map prediction. Traditional encoder-decoder architectures, while designed for cross-modality feature interactions, may not have adequately considered the robustness against noise originating from defective modalities, thereby leading to suboptimal performance in complex scenarios. Inspired by hierarchical human visual systems, we propose the ConTriNet, a robust Confluent Triple-Flow Network employing a "Divide-and-Conquer" strategy. This framework utilizes a unified encoder with specialized decoders, each addressing different subtasks of exploring modality-specific and modality-complementary information for RGB-T SOD, thereby enhancing the final saliency map prediction. Specifically, ConTriNet comprises three flows: two modality-specific flows explore cues from RGB and Thermal modalities, and a third modality-complementary flow integrates cues from both modalities. ConTriNet presents several notable advantages. It incorporates a Modality-induced Feature Modulator (MFM) in the modality-shared union encoder to minimize inter-modality discrepancies and mitigate the impact of defective samples. Additionally, a foundational Residual Atrous Spatial Pyramid Module (RASPM) in the separated flows enlarges the receptive field, allowing for the capture of multi-scale contextual information. Furthermore, a Modality-aware Dynamic Aggregation Module (MDAM) in the modality-complementary flow dynamically aggregates saliency-related cues from both modality-specific flows. Leveraging the proposed parallel triple-flow framework, we further refine saliency maps derived from different flows through a flow-cooperative fusion strategy, yielding a high-quality, full-resolution saliency map for the final prediction. To evaluate the ro

关键词： Decoding Robustness Benchmark testing object detection Noise Lighting Accuracy Visualization Thermal noise Modulation Encoder-Decoder multi-modal fusion RGB- thermal salient object detection

来源：评论

学校读者我要写书评

暂无评论

Cross-Modality object detection Based on DETR

引用

IEEE ACCESS 2025年 13卷 51220-51230页

作者： Huang, Xinyi Ma, Guochun Hangzhou Normal Univ Sch Math Hangzhou 311121 Zhejiang Peoples R China

Cross-modality can integrate complementary information from different modalities to improve the reliability and robustness of object detection effectively. However, compared to processing uni-modality inputs, cross-modality may significantly increase the computational complexity of the model, which will lead to huge memory consumption. In this paper, we propose a novel lightweight Cross-Modality Hybrid Encoder (CHE) that maintains low computational consumption while enhancing the performance of the detection model, which includes two modules: the Attention-based Cross-Modality Feature Interaction (ACFI) module for feature interaction within and between modalities, and the Res-CNN-based Cross-Modality Feature Fusion (RCFF) module for feature association and enhancement. Our method is tested on LLVIP, FLIR, and M3FD datasets. The experiments demonstrates good performance. Our proposed approach outperforms the state-of-the-art methods on mAP with 2.6% on LLVIP, 1.1% on FLIR, and 0.4% on M3FD datasets. Additionally, our method requires only 79.7M parameters, significantly fewer than the 244.6M parameters of the SOTA method Fusion-Mamba, while maintaining high accuracy. This work explores DETR's potential for cross-modality fusion. Our code and models are available at https://***/CVandDetect/Cross-Modality-Detect.

关键词： object detection Feature extraction Transformers Detectors Computational modeling Accuracy Channel hot electron injection Training Standards Semantics Cross-modality feature fusion object detection transformer

来源：评论

学校读者我要写书评

暂无评论

Enhanced YOLOv8 object detection Model for Construction Worker Safety Using Image Transformations

引用

IEEE ACCESS 2025年 13卷 10582-10594页

作者： Seth, Yash Sivagami, M. Vellore Inst Technol Sch Comp Sci & Engn Chennai 600127 India

The rapid growth of Deep Learning techniques plays a vital role in automation of manual work in various areas. One such area for application of new technology is that of Construction Worker Safety. It has thus become imperative to improve existing systems with the new capabilities of technology. This paper discusses a methodology of improving the performance of an existing approach of object detection, YOLOv8. The proposed work comprises of improved training of model and detection of helmet in worker images, using Test Time Augmentation (TTA) based approach. Image Transformations such as Histogram Equalization, Gamma Correction, Gaussian Blurring and Contrast Stretching are applied to augment the dataset by creating more versions of the existing data. This has shown to improve the performance of the model and also generalize better by preventing overfitting. A Test Time Augmentation-based Confidence Thresholding formula (TTACT) is also proposed, to improve the performance of helmet detection.

关键词： Safety Head Deep learning Accuracy Data models Feature extraction Image transformation YOLO Data augmentation Robustness Construction worker safety data augmentation deep learning image transformation object detection test time augmentation YOLOv8

来源：评论

学校读者我要写书评

暂无评论

SpermDet: Structure-Aware Network With Local Context Enhancement and Dual-Path Fusion for object detection in Sperm Images

引用

IEEE TRANSACTIONS ON INSTRUMENTATION AND MEASUREMENT 2025年 74卷

作者： Zhang, Hongyu Hu, Zhujun Huang, Huaying Liu, Shuang Rao, Yunbo Wang, Qifei Ahmad, Naveed Univ Elect Sci & Technol China Sch Informat & Software Engn Chengdu 611731 Peoples R China Southwest Med Univ Dept Reprod Med Affiliated Hosp Luzhou 646000 Peoples R China Google Inc Berkeley CA 94720 USA Prince Sultan Univ Coll Comp & Informat Sci Riyadh 12435 Saudi Arabia

Recently, deep-learning-based object detection models have been used to improve the detection performance in sperm images. However, these models encounter three primary challenges: 1) visual similarity between sperm and background noise;2) neglecting critical local contextual features;and 3) diminishment of tiny sperm features when fusing deep features with an inaccurate foreground. In this article, we propose SpermDet to alleviate these issues. Specifically, we first develop a structure-aware alignment fusion (SAF) module to align and integrate structural information with RGB images, thus improving feature discriminability. Then, we introduce a local context enhancement block (LCEB) to effectively capture the crucial local contextual features of sperm. Furthermore, we design a semantic-aware dual-path fusion (SDF) module that uses both foreground and background information of deep layers to enhance semantic information and preserve detailed sperm features in shallow layers. Finally, we construct a dataset for sperm detection in the testicular biopsy scene, termed as SDTB. It contains 1341 images with 5548 instances, characterized by challenges of tiny sperm size and complex backgrounds. Extensive experiments on the SVIA, VISEM, and the proposed SDTB datasets demonstrate that SpermDet outperforms 24 recent models, achieving mAP(50) scores of 77.7%, 26.3%, and 53.2%, respectively. Compared with the baseline model, SpermDet achieves mAP(50) score improvements of 5.0%, 6.9%, and 9.2% on these datasets while reducing the parameters by 16.3%, which has the potential for application in clinical instruments. The code and dataset are publicly available at: https://***/Hong-yu-Zhang/SpermDet.

关键词： Feature extraction object detection Noise Context modeling Background noise Semantics Visualization Transformers Neck Electronic mail Computer-aided semen analysis (CASA) system feature fusion local context extraction multimodal image fusion object detection

来源：评论

学校读者我要写书评

暂无评论

Large Model-Assisted Federated Learning for object detection of Autonomous Vehicles in Edge

引用

IEEE TRANSACTIONS ON VEHICULAR TECHNOLOGY 2025年第2期74卷 1839-1848页

作者： Behera, Saswat Adhikari, Mainak Menon, Varun G. Khan, Mohammad Ayoub Indian Inst Informat Technol Lucknow Lucknow 226002 India Indian Inst Sci Educ & Res Thiruvananthapuram Thiruvananthapuram 695551 India SCMS Sch Engn & Technol Karukutty 683576 India Univ Bisha Coll Comp & Informat Technol Bisha 67714 Saudi Arabia

The advancement of Autonomous Vehicles (AVs) significantly relies on the integration of Internet-of-Things technology for real-time data processing and decision-making. object detection, a critical component of AVs, necessitates utilizing machine learning techniques. However, this poses challenges including data privacy and the inefficiency of transmitting large volumes of data to a central server. To overcome these challenges, we employ Federated Learning (FL), which allows local devices to train a global model without sharing their raw data. While effective, synchronous federated learning can be time-consuming due to stragglers and often encounter device dropout issues. In contrast, asynchronous federated learning provides faster updates but frequently yields sub-optimal models compared to its synchronous counterpart. Motivated by that, in this paper, we introduce a novel FL framework that combines the strengths of both synchronous and asynchronous methods. Further, by organizing devices into a hierarchical structure, we aim to optimize model convergence while mitigating straggler and dropout problems. Moreover, deploying large models on local edge devices is impractical due to limited computational capabilities, necessitating the adoption of lightweight models to mitigate prolonged training and prediction duration. To evaluate the performance of the proposed method, we conduct a series of experiments using publicly available datasets and subsequently compare the results. Our findings demonstrate that the proposed method significantly enhances the model convergence and performance in object detection for AVs.

关键词： Computational modeling Servers Training Data models Image edge detection Federated learning Autonomous vehicles object detection autonomous vehicles (AVs) federated learning (FL) large model edge computing

来源：评论

学校读者我要写书评

暂无评论

SFFEF-YOLO: Small object detection network based on fine-grained feature extraction and fusion for unmanned aerial images

引用

IMAGE AND VISION COMPUTING 2025年 156卷

作者： Bai, Chenxi Zhang, Kexin Jin, Haozhe Qian, Peng Zhai, Rui Lu, Ke Henan Univ Sch Software Kaifeng 475000 Peoples R China Henan Univ Henan Prov Engn Res Ctr Intelligent Data Proc Kaifeng 475000 Peoples R China Univ Chinese Acad Sci Sch Engn Sci Beijing 100049 Peoples R China

Unmanned aerial vehicles (UAVs) images object detection has emerged as a research hotspot, yet remains significant challenge due to variable target scales and the high proportion of small objects caused by UAVs' diverse altitudes and angles. To address these issues, we propose a novel Small object detection Network Based on Fine-Grained Feature Extraction and Fusion(SFFEF-YOLO). First, we introduce a tiny prediction head to replace the large prediction head, enhancing the detection accuracy for tiny objects while reducing model complexity. Second, we design a Fine-Grained Information Extraction Module (FIEM) to replace standard convolutions. This module improves feature extraction and reduces information loss during downsampling by utilizing multi-branch operations and SPD-Conv. Third, we develop a Multi-Scale Feature Fusion Module (MFFM), which adds an additional skip connection branch based on the bidirectional feature pyramid network (BiFPN) to preserve fine-grained information and improve multi-scale feature fusion. We evaluated SFFEFYOLO on the VisDrone2019-DET and UAVDT datasets. Compared to YOLOv8, experimental results demonstrate that SFFEF-YOLO achieves a 9.9% mAP0.5 improvement on the VisDrone2019-DET dataset and a 3.6% mAP0.5 improvement on the UAVDT dataset.

关键词： Unmanned Aerial Vehicles (UAVs) object detection Small objects Feature extraction Feature fusion

来源：评论

学校读者我要写书评

暂无评论

GARD: A Geometry-Informed and Uncertainty-Aware Baseline Method for Zero-Shot Roadside Monocular object detection

引用

IEEE ROBOTICS AND AUTOMATION LETTERS 2025年第2期10卷 1297-1304页

作者： Peng, Yuru Wang, Beibei Yu, Zijian Zhang, Lu Ji, Jianmin Zhang, Yu Zhang, Yanyong Univ Sci & Technol China Inst Adv Technol Hefei 230026 Peoples R China Hefei Comprehens Natl Sci Ctr Inst Artificial Intelligence Hefei 230026 Peoples R China Hefei Comprehens Natl Sci Ctr Inst Artificial Intelligence Hefei 230026 Peoples R China Univ Sci & Technol China Sch Comp Sci & Technol Hefei 230026 Peoples R China Suzhou Shuzhi Technol Grp Co Ltd Suzhou 215000 Peoples R China

Roadside camera-based perception methods are in high demand for developing efficient vehicle-infrastructure collaborative perception systems. By focusing on object-level depth prediction, we explore the potential benefits of integrating environmental priors into such systems and propose a geometry-based roadside per-object depth estimation algorithm dubbed GARD. The proposed method capitalizes on the inherent geometric properties of the pinhole camera model to derive depth as well as 3D positions for given 2D targets in roadside-view images, alleviating the need for computationally intensive end-to-end learning architectures for monocular 3D detection. Using only a pre-trained 2D detection model, our approach does not require vast amounts of scene-specific training data and shows superior generalization abilities across varying environments and camera setups, making it a practical and cost-effective solution for monocular 3D object detection.

关键词： Three-dimensional displays Cameras Roads Depth measurement object detection Estimation Feature extraction Detectors Deep learning Accuracy Computer vision for automation intelligent transport systems object detection

来源：评论

学校读者我要写书评

暂无评论

SKL-YOLOv8s: an object detection method for coal gangue flow

引用

INTERNATIONAL JOURNAL OF COAL PREPARATION AND UTILIZATION 2025年

作者： Wang, Guoxin Zhu, Jiandong Wang, Shuxia Ru, Hongfang Heilongjiang Univ Sci & Technol Sch Elect & Control Engn Harbin 150022 Heilongjiang Peoples R China Heilongjiang Univ Sci & Technol Heilongjiang Prov Key Lab Opt 3D Measurement & Det Harbin Peoples R China Harbin Univ Sci & Technol Heilongjiang Prov Key Lab Laser Spect Technol & Ap Harbin Peoples R China

Coal gangue sorting is a critical process in the clean production of coal. To address the issues of high computational complexity and deployment challenges in coal gangue object detection tasks, this paper proposes the SML-Neck structure, which incorporates the KernelWarehouse, LSK module, and MPDIoU to enhance YOLOv8s, resulting in the lightweight SKL-YOLOv8s model. Experiments were conducted using a self-constructed dataset comprising 14,790 images of coal gangue. The experimental results demonstrate that the proposed method surpasses YOLOv5s, YOLOv5x, YOLOv7, YOLOv7x, SSD, Faster-RCNN, RT-DETR-l, RT-DETR-x, and YOLOv8s in both accuracy and speed. Compared to the original model, it achieves a 60.9% reduction in computational cost, a 49.5% improvement in inference speed, and a decrease in information transmission loss from 2.6 to 0.6. The GradGAM visualization algorithm reveals that the improved model concentrates its attention more effectively on the entire target, enhancing the network's ability to distinguish between objects and background. This method contributes to advancing clean coal development and provides a significant reference for coal gangue sorting based on object detection.

关键词： Coal gangue object detection YOLOv8 lightweighting real-time sorting

来源：评论

学校读者我要写书评

暂无评论

SGFNet: Structure-Guided Few-Shot object detection

引用

IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS FOR VIDEO TECHNOLOGY 2025年第4期35卷 3209-3221页

作者： Ma, Jingkai Bai, Shuang Beijing Jiaotong Univ Sch Elect & Informat Engn Beijing 100044 Peoples R China

Few-shot object detection (FSOD) focuses on detecting objects of novel classes with only a small number of annotated samples. Due to the limited number of new class samples and the presence of intra-class variance, current FSOD methods struggle to acquire sufficient discriminative information to represent the corresponding class, thus restricting the performance of FSOD. To address this issue, we propose a Structure-Guided Few-shot object detection (SGFNet) method that utilizes the structural information of targets to provide richer discriminative information. Specifically, we first design a Multi-Frequency Structural Feature (MFSF) module, where the highly discriminative structural information of objects in images is extracted and used to enhance the discriminativeness of the features of the target. Based on the MFSF, we then propose a Saliency Information Enhancement (SIE) module that utilizes saliency information to enhance the object-related structural features while suppressing background interference. In addition, we present a novel Soft Cosine Classifier (SCC) based on soft cosine similarity to extract consistent discriminative information between the support and query features for distinguishing targets. Extensive experiments on PASCAL VOC and MS COCO demonstrate that our method significantly outperforms a strong baseline (up to 13.8%) and previous state-of-the-art methods (4.8% in average).

关键词： Feature extraction object detection Data mining Interference Few shot learning Visualization Training Frequency-domain analysis Filtering Correlation Few-shot object detection structural information saliency information soft cosine similarity

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：